Homeowner preferences after September 11th, a microdata approach

Abstract

The existence of homeowner preferences-specifically homeowner preferences for neighborsis fundamental to economic models of sorting. This paper investigates whether or not the terrorist attacks of September 11, 2001 (9/11) impacted local preferences for Arab neighbors. We test for changes in preferences using a differences-indifferences approach in a hedonic pricing model. Relative to sales before 9/11, we find properties within 0.1 miles of an Arab homeowner sold at a 1.4% discount in the 180 days after 9/11. The results are robust to a number of specifications including time horizon, event date, distance, time, alternative ethnic groups, and the presence of nearby mosques. Previous research has shown price effects at neighborhood levels but has not identified effects at the micro or individual property level, and for good reason: most transaction level data sets do not include ethnic identifiers. Applying methods from the machine learning and biostatistics literature, we develop a binomial classifier using a supervised learning algorithm and identify Arab homeowners based on the name of the buyer. We train the binomial classifier using names from Summer Olympic Rosters for 221 countries during the years 1948-2012. We demonstrate the flexibility of our methodology and perform an interesting counterfactual by identifying Hispanic and Asian homeowners in the data; unlike the statistically significant results for Arab homeowners, we find no meaningful results for Hispanic and Asian homeowners following 9/11.

Are historic districts a backdoor for segregation? Yes and no

“…It is also common in economic research. Nowak and Sayago-Gomez (2018) use the Olympic rosters as the training data set and develop a binomial ethnicity classifier based on names to predict whether home buyers are Arabic or not. Humphreys et al (2019) use the same method to predict whether home buyers are Chinese or Korean.…”

“…They use these datasets to predict race and ethnicity based on the first and last name, or just the last name. Note that they did not directly use a complete name as in Nowak and Sayago-Gomez (2018) and Humphreys et al (2019); rather, they split the strings into two character chunks (bi-chars) similar to that in Ambekar et al (2009) and Xu (2019). For example, a name Smith becomes Sm, mi, it, and th.…”

“…On the one hand, these papers use slightly different methods: Nowak and Sayago‐Gomez (2018) and Humphreys et al. (2019) use a first or last name as the token to train and predict the ethnicity, and they use a logistic model and Lasso (least absolute shrinkage and selection operator) approach.…”

“…However, Sood and Laohaprapanon (2018) are trying to predict the race of English names within the U.S. into four categories: White, Black, Asian, and Hispanic. Xu (2019) is predicting ethnicity for people from Poland, Germany, and Russia, who tend to have more unique strings in their names; so is the case for the Arabic setting in Nowak and Sayago‐Gomez (2018) and the Chinese and Korean setting in Humphreys et al. (2019).…”

Using machine learning to identify spatial market segments. A reproducible study of major Spanish markets

“…The information obtained is then used as an input in econometric hedonic models. For example, Humphreys et al (2019) and Nowak and Sayago-Gomez (2018) used machine learning classifiers to ethnically profile buyers and sellers based on last names to understand whether potential cultural biases and/or discrimination issues exist in property transactions. In other research, machine learning algorithms replace the conventional hedonic price model (Hu et al, 2019; Yoo et al, 2012; Füss and Koller, 2016).…”